2026-06-09 20:03:29
Welcome to the Premium edition of How They Make Money.
Over 300,000 subscribers turn to us for business and investment insights.
In case you missed it:
That’s more than what the largest IPO in history will raise.
SpaceX is about to price the largest initial public offering ever at $75 billion, valuing the company at ~$1.8 trillion. But Alphabet, the parent company of Google, quietly topped that last week by selling new shares — upsizing a planned $80 billion raise to $85 billion after demand poured in. Berkshire Hathaway anchored it with a $10 billion investment, buying shares at a 6–8% discount.
Here’s the paradox. Alphabet sits on well over $120 billion in cash and short-term investments on its balance sheet, just generated $165 billion in operating cash flow last year, and can borrow at rates few companies on earth can match. Firms in that position almost never sell equity. Instead, they use their cash flow to buy back stock. Diluting your own shareholders is the most expensive money there is, a slice of every future dollar of profit, forever.
So why do it? Because the AI buildout has grown so large it requires every available tool. Alphabet expects to spend $190 billion in 2026 alone, and that figure could increase by year-end.
Every company racing to build compute now stands in front of three doors:
Pay with the cash you already generate.
Borrow it (issue debt).
Sell a piece of the company (issue equity).
Each door sends a different signal about how confident you are and how much risk you are willing to take (more on this in a minute).
Today, we are looking at the three ways Big Tech is currently funding the trillion-dollar AI buildout. We’ll review which ones are good or bad for investors and what it says about where we are in this market cycle.
💵 Self-funding
🏦 Debt issuance
📉 Selling more stock
The cleanest way to fund a data center is to pay for it with money the business already makes. No interest, no dilution, no banker fees. For most of this cycle, that’s exactly what Big Tech did.
If you need a reminder, free cash flow (FCF) is the cash generated from the business, less capital expenditures (funds to buy, upgrade, or maintain physical assets). As long as FCF is positive, you are effectively self-funding your business ventures.
Free Cash Flow = [Cash from Operations – Capital Expenditures]Amazon is the textbook case for how far that model stretches. The company is on track for roughly $200 billion in capital expenditures in 2026, up from a trailing-twelve-month figure of $147 billion (+67% Y/Y).
Meanwhile, trailing-12-month operating cash flow grew to $149 billion (+30% Y/Y) at the end of March. It’s a staggering number that, a few years ago, would have funded the whole thing with room to spare. It no longer does.
Amazon is now spending essentially every dollar it generates. It was already the case in 2021 and 2022, when the company invested in fulfillment to meet pandemic-driven demand. As a result, FCF turned slightly negative in 2021 and 2022. FCF recovered to a record $38 billion in 2024 before the AI buildout cut it to $11 billion in 2025
At the end of Q1 2026, trailing FCF worsened to just ~$1 billion.
“The faster AWS grows, the more short-term CapEx we’ll spend.”
That was CEO Andy Jassy on the Q1 call, and it captures the squeeze. AWS reaccelerated to +28% Y/Y, its fastest in nearly four years, but Amazon has to lay out cash for land, power, chips, and servers 6 to 24 months before it can bill a customer. Growth and free cash flow are pulling in opposite directions.
Self-funding is a position of strength: you don’t need anyone’s permission to build. The catch is that it caps your spend at what the business produces today. The AI bill is now testing that ceiling everywhere. Microsoft, Meta, and Alphabet still fund most of their CapEx from existing operations. The open question is what happens when one great business can no longer feed an even hungrier one.
Takeaway: Free cash flow is the cheapest capital and the proudest signal — it says the core business can pay its own way. But when CapEx consumes more than 100% of operating cash flow, the wallet is empty, and the next dollar has to come from somewhere else.
2026-06-06 22:02:53
Welcome to the Saturday PRO edition of How They Make Money.
Over 300,000 subscribers turn to us for business and investment insights.
In case you missed it:
📊 Monthly reports: 200+ companies visualized.
📩 Tuesday articles: Exclusive deep dives and insights.
📚 Access to our archive: Hundreds of business breakdowns.
📩 Saturday PRO reports: Timely insights on the latest earnings.
Today at a glance:
📈 Broadcom: AI Bookings Triple Shipments
🖥️ HPE: AI Demand Pulls Forward
🧑⚕️ Veeva: Falcon Targets Agentic Labor
🌐 Samsara: Operational AI Scales
🔷 Rubrik: Agentic Cyber Resilience
🧘🏻 Lululemon: Turnaround Stalls
🛠️ GitLab: AI Pivot With Layoffs
Broadcom's Q2 revenue rose 48% Y/Y to $22.2 billion ($70 million beat), with non-GAAP EPS of $2.44 ($0.04 beat).
Semiconductor solutions as a whole grew 79% to $15 billion, driven by AI semiconductor revenue, which jumped 143% Y/Y to $10.8 billion.
Infrastructure software grew 9% to $7.2 billion.
Operating margin hit a record 49%, and free cash flow reached $10.3 billion (46% of revenue). Shares still fell nearly 15% post-earnings, giving back most of the $270 billion in market cap added over the prior five sessions.
Broadcom is seen as a major beneficiary of Google's announcement earlier this week that it is raising ~$85 billion in equity to fund its AI infrastructure buildout. Google is one of Broadcom's six core custom XPU customers.
For Q2, Broadcom’s headline metric was order velocity. AI semiconductor bookings exceeded $30 billion, roughly three times what the company shipped, with CEO Hock Tan saying “our visibility runs all the way to 2028.” AI networking accounted for nearly 40% of AI revenue, though Tan expects this to normalize toward 30% as custom XPUs ramp.
Expectations were sky-high ahead of the print:
Q3 AI semiconductor guidance of $16 billion implies a 200% Y/Y growth, but it missed the $17.2 billion buy-side estimate.
Tan implied full-year FY26 AI revenue of around $56 billion, which was short of the $57.6 billion consensus.
The $100 billion FY27 AI target was reiterated during the call, but not raised.
Multi-year customer commitments are real but back-end loaded into late FY27 and FY28.
Broadcom announced major customer-specific deployments:
1.3 GW for OpenAI in 2027 (part of a 10 GW deal through 2029).
3 GW for Meta through the end of 2028 (with the first 1 GW starting in H2 2027).
A $35 billion debt financing deal with Apollo and Blackstone to fund Anthropic's purchase of Google TPU chips that Broadcom helped develop, with Broadcom backstopping the largest portions of the transaction. The arrangement is part of a broader AI XPU Platform targeting 20+ GW of LLM compute capacity through 2028.
Broadcom guided Q3 revenue to $29.4 billion (vs. $28.47 billion consensus), with semiconductor revenue at $20.5 billion (up 124% Y/Y).
The sell-off was primarily about Broadcom failing to raise its $100 billion FY27 AI revenue target, compounded by Tan acknowledging that Google will diversify its TPU supply. Investors are left wondering whether management is staying cautious to set up future beats or whether the strength in bookings is already priced in.
2026-06-05 20:02:35
Welcome to the Free edition of How They Make Money.
Over 300,000 subscribers turn to us for business and investment insights
In case you missed it:
Palo Alto Networks and CrowdStrike are the two largest pure-play cybersecurity companies in the world, and they reported a day apart this week with almost the same script: strong results, higher guidance, and a stock that fell anyway.
Both stocks had rallied more than 50% into earnings as investors embraced the same thesis: AI is becoming a tailwind for cybersecurity.
The trigger was Anthropic’s Mythos, the model judged too dangerous to release widely, which sent enterprises scrambling to reassess their defenses.
The numbers suggest the AI security thesis is real. The problem is that the investors already knew it.
Today at a glance:
☁️ Palo Alto: Buying the AI Security Era
🦅 CrowdStrike: The Mythos Moment
Palo Alto delivered a strong quarter, but the headline numbers were messy.
Revenue grew 31% Y/Y (or 14% organically) to $3.0 billion ($60 million beat), and non-GAAP EPS landed at $0.85 ($0.05 beat).
The company also reported an unusual $183 million operating loss, but that was mostly deal-related noise from the recent $25 billion CyberArk acquisition and the Chronosphere deal. New acquisitions added roughly $280 million of intangible amortization and $198 million of deal costs this quarter alone.
Cash flow told a cleaner story. Operating cash flow rose 39% Y/Y to $871 million, while adjusted free cash flow hit a record $910 million, up 57% Y/Y. The trailing-12-month free cash flow margin reached 38.5%, up 4.3 percentage points Y/Y.
Management leaned into the early-2026 SaaSpocalypse selloff, spending $1 billion on 6.8 million shares at an average of ~$148 — nearly half today’s price. When you’re that sure the panic is wrong, you buy. That’s also why we built a larger PANW position at the time in App Economy Portfolio.
Did Palo Alto buy its way into the AI security era at the right price?
So far, the answer looks encouraging.
Strip out the acquisition noise, and the organic numbers look healthy:
Next-Gen Security (NGS) ARR: $8.1 billion, up 28% organically.
Remaining performance obligation: $18.4 billion, up 22% organically.
Organic growth matters because it shows that the CyberArk deal did not mask a slowdown. The forward-looking metrics show Palo Alto is still signing large customers, expanding its backlog, and converting more of the business to recurring revenue.
CEO Nikesh Arora framed AI as a forcing function. Frontier models are making companies rethink their defenses, especially as AI agents gain access to credentials, data, and internal systems.
The demand signal is already showing up. Roughly 1,000 companies reached out over two months to reassess their cyber posture, and the new Unit 42 Frontier AI Defense offering generated 800+ customer meetings in six weeks.
Palo Alto is also turning the threat into a tool. With early access to frontier models, the company says it can compress a year’s worth of penetration testing into under three weeks.
But Arora’s most important point was more grounded: AI still makes mistakes. Frontier models can hit error rates near 25% and fail at the last mile of complexity. In cybersecurity, that gap matters. Attackers only need to win once, and one wrong enforcement call can take down a production network. In short, AI expands the attack surface, but customers still need a trusted platform to manage the response.
Palo Alto’s pitch is to replace a dozen point products with one platform, trading upfront discounts for stickier, multi-year commitments.
The strategy is working:
110 net new platformizations this quarter, lifting the total to ~2,280.
120% net retention with single-digit churn. That’s exactly what justifies large onboarding discounts.
Firewall bookings rose across appliances (+40% Y/Y, the strongest hardware quarter in a decade) and software (+25%).
The legacy firewall business still matters, but the mix is changing. 46% of trailing-12-month product revenue is now recurring software, up from 22% three years ago. That transition is the key. Palo Alto is using its firewall footprint to pull customers into a broader subscription platform.
The acquisitions are the clearest sign of where Palo Alto thinks cybersecurity is going. The company is trying to own the full stack before the market fully forms.
Identity: CyberArk, now rebranded Idira, gives Palo Alto a stronger identity layer for humans, machines, and AI agents. As agents spread, credentials become one of the fastest-growing attack surfaces.
Security operations: Cortex XSIAM crossed $600 million ARR, up roughly 100% Y/Y, with more than 740 customers as it displaces legacy SIEM tools.
AI runtime: Prisma AIRS topped 300 customers, tripling sequentially, with a $20 million-plus enterprise deal already signed.
Management raised the full-year outlook for the third straight quarter:
FY26 revenue guidance increased to $11.42 billion ($60 million raise), while non-GAAP EPS moved to $3.77–$3.79 (up from $3.65–$3.70).
Longer term, management still expects a 40% free cash flow margin by FY28 and $20 billion in NGS ARR by FY30. CFO Dipak Golechha also flagged rising memory and storage costs, with a 10% increase in hardware prices now baked into the guide.
Takeaway: Palo Alto’s AI security pitch is credible because the organic engine is still strong. The CyberArk deal adds complexity, but it also gives Palo Alto a better shot at owning identity in the agent era. Arora has earned the benefit of the doubt. The risk is valuation: after a nearly 50% rally this year and a stock near 62x forward EBITDA, investors are already pricing in a clean integration.
CrowdStrike delivered the cleanest validation of the AI thesis yet: Q1-record net new ARR and higher guidance on every line.
After surging over 60% in the month leading up to the print, the bar was set for a flawless report, and a merely excellent one wasn’t enough. The stock fell more than 10%.
Revenue grew 26% Y/Y to $1.4 billion ($30 million beat), the fourth straight quarter of acceleration, and non-GAAP EPS landed at $1.10 ($0.03 beat). Margins improved across the board, with the operating loss narrowing to $31 million (from $125 million a year ago).
Free cash flow hit a record $468 million, a 34% margin and an all-time high, lifting CrowdStrike’s Rule of 40 score to 59. Stock-based compensation remains a significant factor, accounting for 21% of revenue and showing little improvement over time.
The recurring-revenue engine is the main KPI:
Ending ARR rose 24% Y/Y to $5.51 billion, well past the halfway mark toward the $10 billion target set at its 2023 Fal.Con briefing.
Net new ARR, the quarter’s fresh recurring revenue and the truest read on momentum, hit a Q1-record $256 million (+32% Y/Y). Q1 is the weakest quarter seasonally, making the numbers all the more impressive.
The thesis CEO George Kurtz has pushed for a year is now the customer’s thesis. Enterprises increasingly treat CrowdStrike as AI security infrastructure rather than a line item. Post-Mythos, executive-level inquiries surged, and the company stood up Project QuiltWorks, a coalition that helps large customers find and fix vulnerabilities in real time.
It also created a Chief AI and Autonomous Systems Officer role, poaching Bartley Richardson from NVIDIA. It illustrates where CrowdStrike thinks the platform will head over the next decade.
With Falcon Flex, customers commit a security budget upfront and then spend it across modules over time. It remains the consolidation flywheel. The company added 300+ Flex accounts, and Flex ARR nearly doubled to $1.9 billion (more than a third of overall ARR).
Customers who burn through their Flex budget early come back and commit more.
480 ReFlex customers (~25% of Flex accounts) lifted ARR by an average of 26%.
130+ have ‘reflexed’ more than once, for an average 51% uplift over their original commitment.
CrowdStrike is racing to own the attack surfaces AI is creating.
AIDR (AI Detection and Response): Ending ARR grew +250% sequentially, with a Q2 pipeline above $50 million. Kurtz called the ramp the fastest of his career and pegged AIDR as a market potentially larger than endpoint detection itself, as AI agents multiply.
Next-Gen SIEM: Ending ARR topped $600 million, and combined with cloud and identity now exceeds $2 billion.
FalconShield: ARR nearly quadrupled Y/Y as customers scramble to secure AI agents operating inside their SaaS apps.
Management raised the full year on every line. FY27 ARR is now expected to grow +25% Y/Y. Net new ARR guidance rose by more than $50 million, which Kurtz framed as the AI tailwind in action. The Q2 guidance came just in line, though.
Takeaway: AI is turning cybersecurity from a defensive budget line into a strategic priority. Palo Alto is buying its way deeper into identity and agent security, while CrowdStrike is expanding Falcon into AI-native infrastructure. The thesis is working. The stocks are the hard part. After massive rallies, expectations have caught up with the story, and the valuations leave little room for error.
That’s it for today!
Happy investing!
Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Save 15% with this link.
Disclosure: I own PANW and CRWD in App Economy Portfolio. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members.
Author's Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization's views.
2026-06-02 20:03:38
Welcome to the Premium edition of How They Make Money.
Over 300,000 subscribers turn to us for business and investment insights.
🔥 The May report is here, with over 100 businesses visualized!
All the key earnings from the past month in one report.
✔️ Cut through the noise with clear, concise financial snapshots.
✔️ See revenue trends, profit margins, and key takeaways instantly.
Download the full report below or log in to your account.
Here’s a sneak peek. 👀
What to expect in our monthly report?
New IPO: SpaceX
Chip Design: NVIDIA, AMD, ARM.
Infrastructure: Arista, Cisco, CoreWeave, Dell.
Hardware: HP, Lenovo.
Crypto: Circle, Coinbase.
Gig Economy: Uber, DoorDash, Airbnb, Grab, Instacart.
Payments: Intuit, PayPal, Block, Nu, Klarna, Affirm, Toast.
Global Commerce: MercadoLibre, Alibaba, PDD, Sea, Shopify.
Data & AI: Palantir, Snowflake, Datadog, Elastic.
Cybersecurity: Cloudflare, Zscaler, Fortinet.
Productivity: HubSpot, Zoom, Monday.
Software Stack: Salesforce, Workday, Atlassian, Figma.
Gaming: Tencent, NetEase, Nintendo, Sony, Take-Two.
Social & Ads: Reddit, Snap, The Trade Desk, Applovin.
Streaming: Disney, Warner, Paramount.
Sports betting: DraftKings, Flutter.
Retail & Apparel: Walmart, Target, Best Buy, Amer Sports, On.
Travel: Airbnb, Expedia, Marriott, Tripadvisor.
Restaurants: McDonald’s, RBI.
FMCG: Hershey, Kraft.
2026-05-30 22:02:13
Welcome to the Saturday PRO edition of How They Make Money.
Over 300,000 subscribers turn to us for business and investment insights.
In case you missed it:
📊 Monthly reports: 200+ companies visualized.
📩 Tuesday articles: Exclusive deep dives and insights.
📚 Access to our archive: Hundreds of business breakdowns.
📩 Saturday PRO reports: Timely insights on the latest earnings.
Today at a glance:
🛒 Costco: Fuel Powers A Record Quarter
💻 Dell: AI Demand Breaks The Model
📦 PDD: The Pivot Bites
🧠 Synopsys: Elliott Joins The Board
🏗️ Autodesk: Buying Its Way Into Operations
☁️ Zscaler: Guidance Spooks Investors
🌱 MongoDB: Atlas Holds The Line
🔐 Okta: Identity For The Agent Era
🖨️ HP: Memory Cost Bites
🛒 Best Buy: Sales Turn Positive Again
☁️ Nutanix: Supply Headwinds
🤖 UiPath: Agentic AI Pays Off
🔍 Elastic: AI Drives The Beat
FROM OUR PARTNERS
That’s what NVIDIA’s CEO Jensen Huang said about robotics. And stock market expert Ark Invest says it could be a $24 trillion global opportunity.
So it’s a big deal that NVIDIA chose to help Miso Robotics perfect the next generation of restaurant-kitchen AI robots.
Miso’s Flippy Fry Station robot can boost restaurant profits by up to 3X. In fact, it’s already working in real kitchens, with 5m+ baskets fried for brands like White Castle. Now, with big-name customers like Jersey Mike’s, Cinnabon, and Haagen-Dazs just added, Miso is scaling fast.
With 100,000+ US fast-food locations in need, that’s a $4B/year revenue opportunity in the US alone. And Flippy is just one piece of Miso’s platform.
Industry powerhouse Ecolab already invested. Join them as an early-stage Miso shareholder today.
This is a paid advertisement for Miso Robotics’ Regulation A offering. Please read the offering circular at invest.misorobotics.com
Costco’s Q3 FY26 (May quarter) revenue rose 12% Y/Y to $70.5 billion ($0.9 billion beat), and EPS rose 15% Y/Y to $4.93 ($0.01 beat). Comparable sales blew past expectations at 9.8% (vs. 7.8% consensus), outpacing the recent prints from Walmart, Target, and Best Buy.
Adjusting for gas and FX, core comps still grew 6.6% worldwide, with traffic up 2.4% and average ticket up 4.2%. The real standout was gasoline. Costco posted all-time volume records in every four-week period of the quarter, and ancillary comps surged in the high 20% range as fuel-price-sensitive members drove first-time visits to the pump. Costco usually sells gasoline 20–50 cents per gallon below local stations.
Digital is quietly building momentum:
Digitally enabled comp sales grew 20.8%, with personalized recommendations alone generating nearly $5 billion in e-commerce sales.
Conversion rates on those recommendations ran triple the site average.
AI-sourced traffic posted triple-digit growth and the highest conversion rate of any channel, though absolute volume is still small.
Shares dropped nearly 5% after earnings, a sign that 47x forward earnings now requires more than stellar execution to move the stock.
Membership fee income reached $1.37 billion (up 11%), executive memberships hit 41.2 million (up 10%), and the US/Canada renewal rate ticked up 10 basis points to 92.2%. But total paid member growth slowed to 4%, which CFO Gary Millerchip explicitly called “a more normal rate,” with no new country launch to juice sign-ups.
Gross margin compressed by 21 basis points due to softer margins across fresh foods and packaged groceries. Costco also made deliberate price investments.
The tariff story is now the swing factor. Costco has begun submitting Section 301 refund claims to US Customs, with processing expected to take 2 to 3 months, and member returns contingent on how the legal timeline resolves. Management also trimmed FY26 warehouse openings to 26 (from 28), pushing two into FY27.
The question for the coming quarters is whether the gasoline-driven traffic halo will translate into stickier basket growth once fuel comps normalize.
Dell shares just surged nearly 30% post-earnings, and they have now more than tripled so far this year.
The legacy hardware company most investors wrote off years ago is now one of the best-performing stocks of 2026. The numbers behind the story are even more striking than the chart.
2026-05-29 20:03:30
Welcome to the Free edition of How They Make Money.
Over 300,000 subscribers turn to us for business and investment insights.
In case you missed it:
AI is forcing investors to rethink which software models deserve a premium.
Two enterprise software giants reported earnings on the same evening this week.
Snowflake surged 38% after hours.
Salesforce kept sliding, extending a painful year-to-date decline.
The market is drawing a new line in software between companies that get paid more when AI works harder and companies built around seat licenses for humans who may no longer need as many seats. Consumption businesses can grow as usage rises. Traditional SaaS companies must prove AI can replace seat expansion rather than erode it.
Today at a glance:
☁️ Salesforce: AI Metrics Bury the Lede
❄️ Snowflake: The Clean Sheet Advantage
FROM OUR PARTNERS
Most people brushed off Apple’s new Starlink integration for iPhones.
Mode Mobile saw something bigger: billions of new users suddenly within reach.
Mode’s EarnOS already reaches 490M+ users that have earned and saved over $1B, and that’s before global satellite coverage. With SpaceX eliminating “dead zones,” Mode’s earning technology can now reach billions more in unbanked and rural populations worldwide.
Their global expansion is perfectly timed, and investors like you still have a chance to invest in their pre-IPO offering at $0.50/share until the end of the day.
With their recent 32,481% revenue growth and newly reserved Nasdaq ticker, Mode is edging closer to a potential IPO. Last chance to lock in current pricing – share price changes at 5:00 PM ET.
This is a paid advertisement for Mode Mobile Regulation A+ offering. Please read the offering circular at https://invest.modemobile.com/
*Mode Mobile recently received their ticker reservation with Nasdaq ($MODE), indicating an intent to IPO in the next 24 months. An intent to IPO is no guarantee that an actual IPO will occur.
*Mode cumulative revenue includes full year revenue of businesses acquired in 2025.
Salesforce buried its earnings announcement under an avalanche of AI metrics. Tokens processed. ‘Agentic Work Units’ delivered. Records ingested. Big numbers, but still no clear consumption revenue line.
Agentforce ARR crossed the $1 billion mark in roughly 18 months, making it one of the fastest ramps in enterprise software history. But management slightly raised the low end of full-year revenue guidance.
The market has not given the company the benefit of the doubt, with the stock already cut in half from its previous peak. That probably explains why Salesforce announced the largest accelerated share repurchase in company history, valued at $25 billion.
The number that matters to shareholders is harder to find. Strip out Informatica, acquired for $8 billion in November, and organic revenue grew roughly 9% Y/Y. Nothing to write home about.
There’s a lingering disconnect. Salesforce sells the apps that AI agents are supposed to operate on top of, and the market is no longer sure whether that makes the company indispensable or expendable.
Revenue +13% Y/Y (9% organic) to $11.1 billion ($70 million beat).
Operating margin reached 21% (+1pp Y/Y).
Non-GAAP EPS jumped 50% to $3.88 ($0.75 beat).
Current RPO +14% Y/Y $33.6 billion, missing the $34+ billion consensus.
FY27 guidance:
Revenue: $45.9–46.2 billion (low end raised by $100 million).
Operating and free cash flow growth: cut to 4–5%, from 9–10%, due to debt issuance for the buyback.
Q2 revenue guide: $11.27–11.35 billion, with a midpoint below the $11.36 billion consensus.
🤖 Agentforce is real, but still small: Agentforce ARR reached $1.2 billion, up 205% Y/Y, with net new ARR inflecting to roughly $400 million from $260 million and $100 million in the prior two quarters. At most enterprise software companies, that would be the headline. At Salesforce, it’s a footnote because the legacy base is so large.
📉 The second half remains a show-me story: Salesforce needed current (next 12 months) RPO growth above 15% to validate the pitched re-acceleration story. It came in at 13% in constant currency. Agentforce is ramping fast, but the broader growth inflection is still a second-half promise.
🛒 Marketing, Commerce, and Tableau are dragging: Management explicitly cited “ongoing weakness in Marketing and Commerce and increased softness in Tableau bookings and renewals.” These are products Salesforce paid billions to acquire. They are now the drag offsetting Agentforce momentum. The product portfolio looks more fragmented every quarter.
💸 The $25 billion buyback came with a financing cost: Salesforce used debt to accelerate its repurchases, deploying half of its $50 billion authorization at once. Diluted share count fell 10% Y/Y, but higher interest expense led management to cut operating and free cash flow growth guidance. Borrowing to buy back stock can be smart if the shares are cheap and the business re-accelerates. If growth keeps slowing, it adds leverage without solving the core growth problem.
🔍 Usage metrics dazzle but monetization is unclear: Salesforce disclosed 28.6 trillion tokens processed (+152% Q/Q), 3.8 billion Agentic Work Units (+111% Q/Q), and 52 trillion records ingested into Data 360. These are strong signs of adoption, but they still leave investors wondering how much of this usage will turn into durable revenue, and on what terms.
This was the first quarter under Salesforce’s new disclosure framework.
The old six-cloud breakdown collapsed into two buckets:
Applications: +9% Y/Y or +7% Y/Y in constant currency.
Infrastructure & Data: +25% Y/Y or +23% Y/Y in constant currency.
Management says Agentforce is now embedded across every application, so the old cloud-level view no longer reflects how customers buy. That’s defensible. It’s also convenient.
Marketing Cloud weakness and Tableau softness used to be easier to track. Now they show up as call commentary while Informatica helps lift the broader Infrastructure & Data bucket.
Less granularity makes the portfolio look cleaner, just as investors need more visibility.
The most revealing exchange on the call was Goldman’s Gabriela Borges asking how Salesforce will charge for Headless 360, a new offering that lets external AI agents, including Anthropic’s Claude Code, access data stored in Salesforce apps.
Headless 360 is Salesforce’s response to the agentic threat. If agents are going to access customer data anyway, sell them the sanctioned door.
There are already signs of demand. Wedbush noted that Anthropic is one of Salesforce’s largest customers, with usage reportedly rising fivefold in Q1 due to Headless 360.
The problem is pricing. Management did not give a clear answer. President Miguel Milano said Salesforce would “work together with our customers and partners to find a fair way to monetize” the product. That answer is the whole problem in one sentence.
Usage is rising, but Salesforce has not yet explained how Headless 360 becomes a durable revenue stream. The risk is what Borges called “value abstraction”: Salesforce becomes the data store behind the agent instead of the interface where work happens.
That may still be valuable, but it requires a different pricing model. Salesforce has not fully figured it out yet.
That pricing challenge is harder because Salesforce carries so much legacy surface area. Sales Cloud, Service Cloud, Marketing Cloud, Commerce Cloud, Tableau, MuleSoft, Slack, and Informatica all come with different renewal cycles, pricing models, and competitive threats. Agentforce has to lift the entire portfolio at once. That may work over time, but the bull case needs proof faster than the product cycle can deliver it.
Takeaway: The issue is no longer whether Agentforce is real. It is whether AI makes Salesforce more essential or turns it into the data layer behind someone else’s agent. Until cRPO re-accelerates and Headless 360 has a clear pricing model, the stock might stay in the penalty box.
Snowflake walked into earnings as a consensus AI loser in software. It walked out with a 38% after-hours rally that erased its 20% year-to-date drawdown in a single session.
Product revenue grew +34% Y/Y to $1.33 billion, accelerating from +30% Y/Y in the prior quarter and beating management’s own projection by seven percentage points. Management raised full-year product revenue guidance to $5.84 billion (+31% Y/Y), up from $5.66 billion three months ago. Snowflake also doubled down on AWS with a five-year, $6 billion commitment.
The bear case was that agents would route around the data layer. The print suggests agents need the data layer more than humans ever did.
Revenue +33% Y/Y to $1.39 billion ($70 million beat).
Product revenue +34% Y/Y to $1.33 billion.
Operating loss margin -23% (+19pp Y/Y).
Non-GAAP operating margin: 12% (+3pp Y/Y).
Non-GAAP EPS: $0.39 ($0.07 beat).
Net revenue retention: 126%.
FY27 guidance:
Product revenue: +31% Y/Y $5.84 billion (up from $5.66 billion).
Non-GAAP operating margin: 13.5% (up from 12.5%).
Q2 product revenue guide: $1.415–1.42 billion (+30% Y/Y), ahead of the $1.38 billion consensus.
🔮 Growth outlook gets a real boost: The Q1 beat is only one part of the story. Management raised full-year guidance by $180 million and lifted the implied growth rate from 27% to 31%.
💵 The $1M+ cohort is the real signal: 46 customers crossed the $1M trailing-revenue threshold this quarter, nearly double the 26 a year ago. NRR at 126% confirms existing customers are spending more, not just new logos arriving. The consumption model is working.
🤖 AI products are now monetizing data that doesn’t even live in Snowflake: Cortex Code (CoCo) reached over 7,100 active accounts. Snowflake Intelligence accounts more than doubled quarter over quarter. Crucially, CEO Sridhar Ramaswamy noted that customers are using these tools to access data sitting inside Microsoft, Salesforce, and SAP apps (not just Snowflake’s own databases). That’s a wider moat than the original pitch.
📈 Operating leverage is finally showing up: Operating loss margin is still deep in the red but hit a new high of -23%. Non-GAAP operating margin expanded +3pp Y/Y to 12%, and full-year guidance went from 12.5% to 13.5% — despite carrying a ~150bp headwind from the Observe acquisition.
📉 RPO has a footnote: While RPO surged 38% Y/Y to $9.21 billion, it actually missed the $9.43 billion consensus. Worth tracking, but not the main story. Snowflake is a consumption business, so committed backlog matters less than actual usage.
The headline number is the $6 billion AWS deal over five years.
The deal gives Snowflake heavier access to AWS Graviton chips and custom AI accelerators, helping lower compute costs as AI workloads scale. CFO Brian Robins said the agreement was structured to support AI growth without pressuring gross margins.
That matters because Snowflake is guiding to a 75% product gross margin while AI usage expands. In other words, the company is trying to turn AI demand into consumption growth without sacrificing the economics investors care about.
AWS gets more AI workloads. Snowflake gets cheaper compute, better margins, and a stronger pitch to enterprise customers already building on AWS.
Easy to miss next to the AWS number, Snowflake announced the acquisition of Natoma, an enterprise-grade implementation of the Model Context Protocol (MCP).
Snowflake wants to govern not only what AI agents can read, but what they can do.
MCP is the wiring that lets AI agents securely take actions on behalf of an enterprise. That matters because agents that act consume far more compute than agents that simply retrieve answers.
Snowflake spent a decade becoming the governed home for enterprise data. Natoma extends that governance into the agentic workflow.
Databricks, Microsoft Fabric, and Google BigQuery all want the same role. Snowflake’s bet is that enterprise security and identity, not just model quality, will decide who controls the agentic data layer.
Takeaway: Snowflake’s quarter flipped the AI narrative. Agents are not routing around the data layer. They appear to be increasing the need for governed data, secure workflows, and scalable compute. That is exactly where Snowflake wants to sit. The 35%+ rally reflects how washed-out sentiment had become, but the bigger lesson is structural. For the right software model, AI can turn rising usage into rising revenue.
That’s it for today!
Happy investing!
Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Save 15% with this link.
Disclosure: I own CRM and SNOW in App Economy Portfolio. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members.
Author's Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization's views.